FEA program opinions? 19

[MechElement]

Hello, I'm in the market for a license of SWx, but I would also like a good FEA program to accompany SWx. I've heard about NENastran & COSMOS running with SolidWorks, but it appears everyone has a different opinion about them. I'm a mechanical designer and working in the Middle East. I'm returning to the States for good in August to go back to school and finish my BSME.

Which FEA program works best with SWx?

 

[fkmeyers]

Most commercial FEA packages at the level of Algor, ANSYS, ABAQUS, NEiNastran, etc. can do all the basic stuff well. I have fully evaluated Algor but am not a current user. Garland has used Algor for a while so he is in a better position to comment on it. I have used NEiNastran for the last 3 years and before that have used ABAQUS, Mechanica, I-Deas, and MSC. Here is what I like about NEiNastran and Noran Engineering…

1. Support: They more responsive for support than any other FEA company I have ever dealt with. The support staff is very knowledgeable and they can answer questions that even the experts here with 30+ years playing with Nastran can not. My applications tend to be complicated. Lots of prestressed cables and pipes and they have been very good at supporting me. If can not get the model to run, they ask me to send it over. If I can, I do and they fix it and send it back. If I can not (proprietary, classified, etc.) they ask for details, they create a similar model and show me how they did it. I call and someone answers the phone and is ready to help. I email and someone emails back that day with a response I can use to fix the problem. When I was evaluating Algor I could only get support from a sales person so I could not evaluate that. A friend of mine that used Algor before he came here said their support took days and was often staffed with interns. Just someone else’s comments. I am sure Garland has a better handle on Algor’s support and will give it. I am just saying that I have never seen a company so dedicated to supporting customers. They seem to always be there and very willing to help.

2. Enhancements: They have been very good at putting in enhancements that we have requested (and others we should have requested). Again, we do very specialized types of analysis and have unique requirements. We suggest something and 4 weeks later it is in a beta we can try out. I could not believe this because in the past when dealing with MSC either it got ignored or it took 2 years or more and was not exactly what we needed (and no opportunity to try it our before it was in production). I know we are not the only customers they do this for. Of course not everything that we have asked for they put in immediately. Some of the more complicated things we know they are working on but will take longer. Other requests they say may not be possible in the next release. Overall I am very impressed, again as compared to other companies.

3. Cost: I would never have been able to sell this to management without making a cost justification. When we looked at what MSC costs and figured in maintenance for 3 years NE was about half the price. We know good support will save us time and money so we weighed this heavily. I have been burned in the past with other FEA companies on support. Nothing is more frustrating than not being able to get a model to run when you are up against a deadline. Again see #1 above.

4. Great product: Their product (NEiNastran) is absolutely fantastic. I am able to run huge models in hours that would normally take days with MSC. I can use large parabolic tet meshes and not worry about mesh refinement and get answers fast with accuracy. I can get data quickly in the format I need it in to generate reports. The modeler supports the other CAE tools I use and the Editor is very helpful in making quick changes and performing trade studies. This product has doubled my productivity...simply put!

These are some of my “opinions”. I hope this helps someone. Others I work with share similar feelings.

Garland, I would like to apologize for being offensive in previous postings. That was not my intention. My objective for spending my lunch hour responding here is to help.

 

[GBor]

Frank,

Please consider the hatchet buried...and not in one of us! I echo STRONGLY your statements comparing MSC Nastran to NENastran. MSC is expensive, and, in my opinion, less functional (plane stress)...this is based on a review about a year and a half ago.

Three years ago, I would have agreed with you about the support from Algor. They had a salesman to sell you the software, a different salesman to support the software (BAD Idea!...the support guys had to figure out what the sales guy said to get you to buy the software and the sales guys could say anything) and getting tech support took as much as 3 days!!!! A virtual killer for any company!

A reorganization about a year ago has improved things dramatically. One point of contact who has a designated tech person and a back-up. I generally get a response during my phone call and certainly by the end of the day.

Even with the relationship that I've had with Algor over the years, I recommend NENastran in aerospace applications and for new consultants who will need to interface with a variety of clients that may each need a different "flavor" of FEA package.

I think functionally, NENastran and Algor are comparable. Because of this, if the analysis is for in-house use, I recommend the "low-bidder" (my defense contrator background).

I "grew up" with Algor, which is very different from NENastran. It was difficult for me to follow the logic of FEMAP in building the models and I felt like there was a lot more bookkeeping tracking which properties went with which material which went with which element...I didn't think it was very evident without some digging. What I REALLY liked was the way FEMAP kept the geometry and the mesh separate...a HUGE boost if you anticipate several modifications to your design over time. If you are new to the FEA world, the bookkeeping will be natural.

I am now an independent consultant and work closely with NENastran, COSMOS, and Algor users. I'm less comfortable with COSMOS. NENastran has become comfortable and I agree...it is a fantastic product. Algor's engine and post-processor is fantastic...their pre-processor has improved remarkably over the past year.

There's some of my opinions...sorry to take your lunch hour, Frank. I look forward to our future technical exchanges.

Garland

Garland E. Borowski, PE

 

[jcfoley]

Thats more like it. A star for you both.......

I, for one, really appreciate the time others take to help spread their experience & knowledge. I noticed that nobody really supports COSMOS (after testing DesignStar 4.5, I can sort of understand why). Any other opinions on these packages or expecially ones that have not been brought to the table?

Chris Foley
Midland, TX

 

[GBor]

Thanks, Chris,

Now we feel special

Garland E. Borowski, PE

 

[fkmeyers]

Thanks! We are here to help.

Frank

 

[rkn40]

Maybe these articles wil help you

http://www.machinedesign.com/asp/viewSelectedArticle.asp?strArticleId=55362&strSite=MDSite

http://www.deskeng.com/articles/03/aug/software/main_dig.htm

http://www.designnews.com/article/CA109189.html?text=cosmos

http://www.feadomain.com/General.html

 

[rkn40]

One more piece of information. If you want to just learn FEA you can buy a personal edition of Proe and Pro Mechanica for $250. But be aware this version of Pro/Mechanica has basic statics, normal modes, thermal and contact. It can do a optimization also. Here is the link

http://www.journeyed.com/itemDetail.asp?T1=86214903N

Hope this helps

 

[FrancoisB]

All depend what you want to do like FEA. Me I have Ansys 9.0 and the support of my dealers is realy fanstatic ROI eng each time I have need the help me immediately. I choose Ansys for two reasons. Fisrt I have Ansys professional I do linear and some contact non linear and for the moment this is good for me but in the futur If I need more I will buy what I need for only the different price betwen professionnal and structural for exemple. Second reason is my delear skillness they are strong in FEA. Third Cosmos is not powerfull like ansys. All depend what you do like FEA if is it only for you Cosmos is less expensive but not very strong with assembly. I dont use Pro mechanica and I have Pro-E.

Roi eng is located in Toronto and Montreal. Me I'm closed to Montreal

Bye

 

[doc888]

Does anyone have any experience with CAEFEM FEA package? I have been told that this package has lots of features and low cost.

 

[fkmeyers]

At the risk of sounding negative I can tell you about CAEFEM. It is a one guy operation, support is absolutely terrible, maintenance buys you only some bug fixes, and the product is marginal at best. Little has been done with it for the last 5 years. The one man behind the operation used to work for SRAC (COSMOS) as their VP of product development (mid 1990s). He left and within 6 months had his own FEA solution, CAEFEM. FEMAP is their pre and post and that lists for $7000 so you might as well just get NEiNastran which is only a few thousand more. CAEFEM will not run without FEMAP. It is totally dependent on it. Yes, CAEFEM is low cost but you will need a pre and post and that is where the cost comes in. I have seen CAEFEM in action and I was impressed with it speed at coming up with the wrong answer.

OK. I will say something positive about it. It is fast but so is NEiNastran and COSMOS because they all use the same PCGLSS solvers.

 

[MechElement]

rkn40 (Automotive)
21 Apr 05 17:26
One more piece of information. If you want to just learn FEA you can buy a personal edition of Proe and Pro Mechanica for $250. But be aware this version of Pro/Mechanica has basic statics, normal modes, thermal and contact. It can do a optimization also. Here is the link

http://www.journeyed.com/itemDetail.asp?T1=86214903N

Hope this helps

Thanks for the link!

After researching all of these options, I don't think FEA falls into my budget. This will allow me to step into FEA while recuperating from my workstation & design software purchase.

 

[doc888]

Thanks for the info. I will take it off my list of potential FEA packages.

 

[gurmeet2003]

Could someone who has used both ANSYS and ABAQUS give me their opinion about the two softwares? I am current user of ANSYS STRUCTURAL, but am looking at ABAQUS. I am interested in non-linear analysis, particularly nonlinear material analysis.

Thanks,

Gurmeet

 

[fkmeyers]

I used Pro Mechanica for a while when it was still owned by Rasna. I loved the interface. It certainly was easy to use and very powerful for optimization. If you do go with it and feel the solver is limited (and it is) you can get the NEiNastran solver by itself for $1500 and use it with the built in Nastran interface they have. You can ask Noran about that. I think you can get a free demo as well from them.

 

[mechengr1313]

gurmeet2003,

I have used ANSYS and now use ABAQUS. First of all, ABAQUS is a very powerfull program. The problem with that is, in my opinion, ABAQUS is not as user friendly as what you are use to from ANSYS. The ABAQUS people, I think tend to forget not everyone running FEA is a Phd. The user interface in ANSYS is better because it seems they have spent more money on it. ABAQUS on the other hand seems to be less marketing and more solving hard problems. In the end, it will probably be your personal preference which decides it for you.

Mark

 

[corus]

Since Abaqus introduced the preprocessor CAE and the post processor Viewer it has a very user friendly interface. That can't be said some years ago when Abaqus relied mainly on third party pre and post processors to do the work. There are still some situations where the user may have to manually edit the input file, but these are the exceptional cases rather than the norm. Speaking from personal experience the Abaqus people are extremely helpful even when the question posed is so stupid that it's clear that the the brain has decided to have a long weekend and gone home early.
In general, Ansys appeals to the lower end of the market. Abaqus offers much more than that, and at times you need that.

corus

 

[rtmpxr]

to corus
"In general, Ansys appeals to the lower end of the market."
I disagree with such consideration. I was expecting to read the reasons for such conclusion, but it seems I hoped too much. In fact, it was a surprise to see 55 posts with 5 stars on a question translated like "best fea please!". Such a sublime ignorance didn't receive the right answer/question: what is your problem you want to resolve?. Because there is no "panacea" program to cope with all the problems you encounter. In the nuclear field we build our programs to deal with the highly non-linearity problems. Does this means that ANSYS or ABAQUS are not good ? Of course not, they are excellent programs for what they were build to do.
This is just an opinion, nothing more than that.

 

[GBor]

rtmpxr,

I certainly hope your "bark is worse than your bite". While I agree with your comments regarding the need to support a position as well as your position with regards to Ansys, the original question was "best fea (that works with SolidWorks) please!". This is an offshoot of a thread from another forum. I suggested MechElement send it here so that he received a broader base of opinions, rather than the opinions from the software forum in which he originally posted.

The thread evolved around post 30 or so to become a general "Which FEA program do you like best" thread. As "one of the 5 stars", I would like to defend the thread as being filled with supported opinions and suggest that you read posts from fkmeyers, GBor (me), and others. To a few people with no knowledge of FEA, the posts have been quite valuable , and, quite frankly, after more than 10 years in this business, I continue to learn from these threads.

Be careful how you speak on these forums...one day you may work for one of us.

Garland E. Borowski, PE

 

[fkmeyers]

Well put Garland!

 

[FrancoisB]

In general, Ansys appeals to the lower end of the market. I don't have this opinion why? For 9500$ USD for the basic Ansys I think your lower is very high me. But I never work with abaqus and I never heard someting bad of it. And for finish Ansys had buy too many good program for go down. Mebay my answer is emotive because ''I pay realy much higher''. And for the moment my stall is over!!!!!!!!!

without frustration ;-)

the happy man of Ansys

Bye or Ciao bella